© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 8-RES-JB
    11. 11713
    Prev
    Next
    INDEX
    Explanations

    the word "rather" in sentences

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "rather" indicating a contrast or distinction

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    adium
    -0.69
    clud
    -0.68
    office
    -0.68
    chairs
    -0.67
    rament
    -0.67
    gio
    -0.65
    ajor
    -0.63
    ocene
    -0.62
    seller
    -0.62
     Kard
    -0.61
    POSITIVE LOGITS
     than
    0.86
     importantly
    0.82
    thia
    0.68
    than
    0.68
     tame
    0.66
     ironic
    0.64
     Than
    0.63
    ional
    0.62
    onsense
    0.61
     surpr
    0.61
    Activations Density 0.020%

    No Known Activations